“LLMs also tend to hallucinate, a virtually-unsolvable problem where they authoritatively make incorrect statements that creates horrifying results in generative art and renders them too unreliable for any kind of mission critical work.”
Our reading, the sentence that was checked:
LLM hallucination is a virtually unsolvable problem that makes LLMs too unreliable for any mission-critical work.
Assessment: Unresolved Evidence as of
Where and when Ed Zitron wrote it
Where's Your Ed At, They're Looting The Internet, paragraph 53.
- Said
- Source observed
- live page, full text, public
- Evidence as of
- the cutoff for this assessment
- Assessment recorded
- assessment wyea-the-great-looting-of-the-internet-39a24a-a1
We recorded 21 claims from this post. See every claim from this post.
The wording above is the version this site observed on 2026-10-28. The source may have been edited after 2024-04-15. This page does not claim the wording was captured then.
What the claim commits to
- Kind
- Observation
- Deadline, as the speaker wrote it
- None stated. This site does not add one.
- Deadline state when assessed
- No horizon
- Deadline state on 2026-10-29
- No horizon (the release’s reference date)
From the speaker
- LLMs also tend to hallucinate
- a virtually-unsolvable problem
- too unreliable for any kind of mission critical work
Our interpretation
- The ordinary reading is that confidently false answers prevent dependable use in work where errors can have serious consequences.
- Any kind is broader than autonomous decision-making. The claim does not expressly exclude supervised or externally checked workflows.
- For the ongoing state described on April 15, 2024, use January–March 2024, the latest complete quarter before publication. LLM research has no standard reporting quarter, so this is an explicit temporal convention.
- Virtually unsolvable means resistant to practical mitigation sufficient for dependable use, rather than a requirement of mathematical perfection.
Counts as supported if: Evidence for the target period establishes persistent hallucinations and shows that practical controls cannot provide adequate reliability across the mission-critical uses covered by the claim.
Counts as contradicted if: Evidence for the target period demonstrates a dependable mission-critical LLM workflow under realistic conditions, or practical mitigation sufficient to defeat the asserted broad prohibition.
Retrospective: these criteria were written after the outcome was public.
Current assessment
Was it true for the period described?
Outcome: Unresolved
The record was checked, and the evidence available by the cutoff does not settle it.
Scored on a plain reading. The speaker named no measure for this claim, so it was judged on the reading an informed reader would take, stated below under our interpretation. Both reviews had to find that reading fair.
- Evidence as of
- Deadline state when assessed
- No horizon
- Assessment recorded
- Policy version
- 2026-10-28.7
Rationale
Contemporaneous research supports concern about hallucinations and unsupervised legal use. Earlier research also demonstrates mitigation while acknowledging residual errors. These findings do not settle the stronger assertion that hallucinations are practically unsolvable and prevent every mission-critical use. Scoring the whole statement would require narrowing it to autonomous work or treating any remaining error as disqualifying, neither of which Zitron specified.
Limitations
- This evaluates the state described in early 2024, not the capabilities of later models.
- The quarter boundary is an explicit convention rather than a reporting period named by the speaker.
- The available studies cover selected tasks and models.
Evidence (2 items)
Context only: Dahl, Magesh, Suzgun, and Ho, arXiv,
caution against the rapid and unsupervised integration
Why it matters here: The January assessment supports serious concern about unsupervised legal use. Its publication date dates that assessment, not the underlying experiments; it does not establish the impossibility of every mission-critical workflow.
- Publisher
- Dahl, Magesh, Suzgun, and Ho, arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
Context only: Dhuliawala and colleagues, arXiv,
does not remove them completely
Why it matters here: This earlier study reports mitigation with residual hallucinations. It is background evidence against equating an unsolved problem with an unmitigable one, but supplies neither a target-period deployment result nor proof of mission-critical reliability.
- Publisher
- Dhuliawala and colleagues, arXiv
- Kind of source
- Preprint, not peer reviewed
- Describes events of
- Public from
What would settle it
A target-period evaluation of a real mission-critical workflow, with explicit reliability requirements and oversight conditions, could establish a counterexample. Supporting the universal claim would require substantially broader evidence that practical controls cannot meet those requirements.
Where it was said (1)
Said once in the sources this register covers.
: Where's Your Ed At, They're Looting The Internet
LLMs also tend to hallucinate, a virtually-unsolvable problem where they authoritatively make incorrect statements that creates horrifying results in generative art and renders them too unreliable for any kind of mission critical work.
- Where
- paragraph 53
- Published
- 2024-04-15
- Source observed
- , live page
- Access
- full text, public
History
Assessment versions
- Assessment wyea-the-great-looting-of-the-internet-39a24a-a1: Current. Outcome: Unresolved. Evidence as of 2026-10-28. Public date not recorded.
Recorded events
Assessment publishedBy this site
First assessment published after both reviews and the final-text check passed.
About assessment wyea-the-great-looting-of-the-internet-39a24a-a1.